20 CORA D. REEVES 



the blue-red series of experiments, pp. 25 and 48 the blue 

 plate was illuminated through a 35mm. slit, the red through a slit 

 constantly changed in width during the series. Relations 

 were about as follows for the red : 



Ratio of illumination 



Slit width at centre and border 



35 mm 100 90 



24 mm 100 71 



14 m 100 82 



6 m 100 89 



3 m 100 90 



Im 100 95 



For the blue as 100 at the center to 90 at the border. 



C. Experiments to Test Discrimination by the Formation 

 of a Food Association. — In this section are presented all 

 experiments in which the attempt was made to establish 

 a food association, together with a few experiments on 

 innate response to white-light differences. In the case 

 of each individual fish used the chronological order of the 

 tests has been preserved, so that the past experience of 

 each fish may be known in its relation to any particular 

 series of tests. 



1. Intensity Discrimination of Dace 



To test the effectiveness of differences in intensity of 

 white light in control of behavior, a horned dace or creek 

 chub {Semotilus atromaculatus Mitchill) which had been 

 in the laboratory for two years but which had not been 

 worked with, was placed in the experiment aquarium. 

 The ceiling lamp was lighted. After some hours the fish 

 was closed into the retention compartment, the two Nernst 

 lamps were lighted, and a slit with a 5.0 mm. aperture 

 was placed in one side of the lamphouse, while the oppo- 

 site side remained without a slit. The slit cut off about 

 three-fourths of the total light reaching the stimulus plate 

 upon that side. (See graph, fig. 7.) Both stimulus patches 

 looked bright, but the one without the slit was of dazzling 

 brightness. 



By moving partitions A and B (fig. 6) this fish, called 

 Bu, was brought into the discrimination compartment. 

 On opening the slide door Bu swam up to the stimulus 

 patches. Of the first seven responses, five were to the 



